Skip to content

build-pudl

build-pudl #18

Workflow file for this run

---
name: build-pudl
on:
workflow_dispatch:
push:
tags:
# Run a build when a stable tag is pushed. If a build has already been run
# successfully from the tagged commit, it will skip the build and trigger a deployment
- "v20*"
schedule:
# 6am UTC daily (11pm PDT, 2am EDT)
# But only if there are changes since the last nightly build.
- cron: "0 6 * * *"
env:
GCP_BILLING_PROJECT: ${{ secrets.GCP_BILLING_PROJECT }}
GCS_OUTPUT_BUCKET: gs://builds.catalyst.coop
BATCH_JOB_JSON: batch_job.json
SKIP_BUILD: "false"
GIT_TAG: ""
BUILD_ID: ""
jobs:
build_and_deploy_pudl:
name: Kick off a PUDL build and deployment with Google Batch
runs-on: ubuntu-latest
permissions:
contents: write
id-token: write
steps:
- name: Checkout Repository
uses: actions/checkout@v7
with:
fetch-depth: 0
- name: Skip the build if no changes since the last successful nightly build.
if: ${{ (github.event_name == 'schedule') }}
run: |
CURRENT_COMMIT=$(git rev-parse HEAD)
NIGHTLY_COMMIT=$(git rev-parse origin/nightly)
if [[ "$CURRENT_COMMIT" == "$NIGHTLY_COMMIT" ]]; then
echo "::notice::No changes since last successful nightly build. Skipping."
echo "SKIP_BUILD=true" >> "$GITHUB_ENV"
exit 0
fi
- name: Set git tag and deployment type based on build type
if: ${{ env.SKIP_BUILD != 'true' }}
# Nightly builds and builds triggered by stable tag pushes trigger production
# deployments. Branch builds trigger staging deployments
run: |
BUILD_ID="$(date +%Y-%m-%d-%H%M)-$(git rev-parse --short=9 HEAD)"
echo "BUILD_ID=$BUILD_ID" >> "$GITHUB_ENV"
if [[ ${{ github.event_name }} == "schedule" ]]; then
echo "GIT_TAG=nightly-$(date +%Y-%m-%d)" >> "$GITHUB_ENV"
echo "DEPLOYMENT_ENVIRONMENT=production" >> "$GITHUB_ENV"
elif [[ ${{ github.event_name }} == "workflow_dispatch" ]]; then
echo "GIT_TAG=branch-$BUILD_ID" >> "$GITHUB_ENV"
echo "DEPLOYMENT_ENVIRONMENT=staging" >> "$GITHUB_ENV"
elif [[ ${{ github.event_name }} == "push" ]]; then
echo "GIT_TAG=${{ github.ref_name }}" >> "$GITHUB_ENV"
echo "DEPLOYMENT_ENVIRONMENT=production" >> "$GITHUB_ENV"
fi
- name: Show freshly set envvars
if: ${{ env.SKIP_BUILD != 'true' }}
run: |
echo "GIT_TAG: $GIT_TAG"
echo "BUILD_ID: $BUILD_ID"
echo "DEPLOYMENT_ENVIRONMENT: $DEPLOYMENT_ENVIRONMENT"
- name: Tag build
if: ${{ (env.SKIP_BUILD != 'true') && (github.event_name != 'push') }}
run: |
git config user.email "pudl@catalyst.coop"
git config user.name "pudlbot"
git tag -a -m "$GIT_TAG" "$GIT_TAG" ${{ github.ref_name }}
git push origin "$GIT_TAG"
- name: Compile notification content
run: |
if [[ "${{ env.SKIP_BUILD }}" == 'true' ]]; then
echo "notification_content=:ghost: Skipping nightly build for $(date +%Y-%m-%d)" >> "$GITHUB_ENV"
else
echo "notification_content=:rocket: Launching PUDL Build: ${{ env.BUILD_ID }} (${{ env.GIT_TAG }})" >> "$GITHUB_ENV"
fi
- name: Notify Zulip of build start or skip
if: ${{ always() }}
uses: zulip/github-actions-zulip/send-message@v2
with:
api-key: ${{ secrets.ZULIP_API_KEY }}
email: "build-status-bot@catalyst-cooperative.zulipchat.com"
organization-url: "https://catalyst-cooperative.zulipchat.com"
type: "stream"
to: "pudl-deployments"
topic: "build-deploy-pudl"
content: "${{ env.notification_content }}"
- name: Docker Metadata
id: docker_metadata
if: ${{ env.SKIP_BUILD != 'true' }}
uses: docker/metadata-action@v6
with:
images: catalystcoop/pudl-etl
flavor: |
latest=auto
tags: |
type=raw,value=${{ github.ref_name }}
type=ref,event=tag
- name: Set up Docker Buildx
if: ${{ env.SKIP_BUILD != 'true' }}
uses: docker/setup-buildx-action@v4
- name: Login to DockerHub
if: ${{ (github.event_name != 'pull_request') && (env.SKIP_BUILD != 'true') }}
uses: docker/login-action@v4
with:
username: ${{ secrets.DOCKERHUB_USERNAME }}
password: ${{ secrets.DOCKERHUB_TOKEN }}
- name: Build image and push to Docker Hub
id: docker-build
if: ${{ env.SKIP_BUILD != 'true' }}
uses: docker/build-push-action@v7
with:
context: .
file: builds/Dockerfile
push: ${{ github.event_name != 'pull_request' }}
tags: ${{ steps.docker_metadata.outputs.tags }}
labels: ${{ steps.docker_metadata.outputs.labels }}
# Docker caching disabled -- see docker-build-test.yml.
- id: "auth"
if: ${{ env.SKIP_BUILD != 'true' }}
uses: "google-github-actions/auth@v3"
with:
workload_identity_provider: "projects/345950277072/locations/global/workloadIdentityPools/gh-actions-pool/providers/gh-actions-provider"
service_account: "deploy-pudl-github-action@catalyst-cooperative-pudl.iam.gserviceaccount.com"
# Setup gcloud CLI
- name: Set up Cloud SDK
if: ${{ env.SKIP_BUILD != 'true' }}
uses: google-github-actions/setup-gcloud@v3
# Deploy PUDL image to GCE
# Dagster Postgres connection configured in builds/dagster.yaml - otherwise we get a str for port num
- name: Make GCP Batch config file
if: ${{ env.SKIP_BUILD != 'true' }}
env:
PUDL_GCS_OUTPUT: ${{ env.GCS_OUTPUT_BUCKET }}/${{ env.BUILD_ID }}
run: |-
./devtools/generate_batch_config.py \
--container-image "docker.io/catalystcoop/pudl-etl@${{ steps.docker-build.outputs.digest }}" \
--container-command "pixi" \
--container-arg="run" \
--container-arg="bash" \
--container-arg="./builds/pudl_batch.sh" \
--container-env AWS_ACCESS_KEY_ID=${{ secrets.AWS_ACCESS_KEY_ID }} \
--container-env AWS_DEFAULT_REGION=${{ secrets.AWS_DEFAULT_REGION }} \
--container-env AWS_SECRET_ACCESS_KEY=${{ secrets.AWS_SECRET_ACCESS_KEY }} \
--container-env BUILD_ID=${{ env.BUILD_ID }} \
--container-env BUILD_REF=${{ github.ref_name }} \
--container-env DEPLOYMENT_ENVIRONMENT=${{ env.DEPLOYMENT_ENVIRONMENT }} \
--container-env GCP_BILLING_PROJECT=${{ secrets.GCP_BILLING_PROJECT }} \
--container-env GITHUB_ACTION_TRIGGER=${{ github.event_name }} \
--container-env GIT_TAG=${{ env.GIT_TAG }} \
--container-env OMP_NUM_THREADS=4 \
--container-env PUDL_BOT_PAT=${{ secrets.PUDL_BOT_PAT }} \
--container-env PUDL_GCS_OUTPUT=${{ env.PUDL_GCS_OUTPUT }} \
--container-env DG_NIGHTLY_CONFIG="src/pudl/package_data/settings/dg_nightly.yml" \
--container-env ZULIP_API_KEY=${{ secrets.ZULIP_API_KEY }} \
--container-env ZENODO_SANDBOX_TOKEN_PUBLISH=${{ secrets.ZENODO_SANDBOX_TOKEN_PUBLISH }} \
--container-env ZENODO_TARGET_ENV=${{ (startsWith(github.ref_name, 'v20') && 'production') || 'sandbox' }} \
--container-env ZENODO_TOKEN_UPLOAD=${{ secrets.ZENODO_TOKEN_UPLOAD }} \
--container-env PUDL_FORCE_FERC_TO_SQLITE="true" \
--output ${{ env.BATCH_JOB_JSON }}
# Start the batch job
- name: Kick off batch job
if: ${{ env.SKIP_BUILD != 'true' }}
run: gcloud batch jobs submit build-pudl-${{ env.BUILD_ID }} --config ${{ env.BATCH_JOB_JSON }} --location us-west1
- name: Status emoji
if: ${{ env.SKIP_BUILD != 'true' }}
run: |
if [[ "${{ job.status }}" == "success" ]]; then
echo "emoji=:check:" >> "$GITHUB_ENV"
elif [[ "${{ job.status }}" == "failure" ]]; then
echo "emoji=:x:" >> "$GITHUB_ENV"
else
echo "emoji=:warning:" >> "$GITHUB_ENV"
fi
- name: Notify Zulip of build launch status
if: ${{ env.SKIP_BUILD != 'true' }}
uses: zulip/github-actions-zulip/send-message@v2
with:
api-key: ${{ secrets.ZULIP_API_KEY }}
email: "build-status-bot@catalyst-cooperative.zulipchat.com"
organization-url: "https://catalyst-cooperative.zulipchat.com"
type: "stream"
to: "pudl-deployments"
topic: "build-deploy-pudl"
content: >
${{ env.emoji }} Google Batch job for ${{ env.BUILD_ID }} (${{ env.GIT_TAG }}) launched with status: ${{ job.status }}.